graphite is giving a MemoryError traceback on a server with 10G of free memory
So after a friend (who works for a large ISP) who uses graphite convinced me it was amazing, I set it up. I followed the install doc and got graphite running. Graphite is running with the run-devel-server script, carbon with --debug, and the example client injects metrics every 30 seconds.
Whenever logging into graphite, the main graph is broken and I'm unsure why. In the runserver output, I saw things like:
[09/Nov/2010 17:08:36] "GET /render HTTP/1.1" 500 77007
[09/Nov/2010 17:08:36] "GET /render/
Lucky for me, I've got some django dev experience and a lot of django debugging experience. I mudged settings.py to point at the django-sentry server we already have setup for many of our other django apps. It gave me this traceback:
Request Method: GET
Request URL: http://
100. response = callback(request, *callback_args, **callback_kwargs)
128. image = doImageRender(
283. img = graphClass(
140. self.drawText("No Data", x, y, align='center')
200. extents = self.getExtents
172. F = self.ctx.
Exception Type: MemoryError at /render/
<email address hidden>: graphite $ free -m
total used free shared buffers cached
Mem: 11906 1464 10441 0 194 833
-/+ buffers/cache: 437 11469
Swap: 8189 0 8189
What am I doing wrong to cause this? Thanks!